Contributory Factors to Caregiver Burden in Parkinson's Disease

ABSTRACT

Background

Although there is growing recognition of the relevancy of informal caregivers there is scarce information on the contributory factors of caregiver burden in Parkinson's Disease (PD).

Objective

To identify the main associated factors to caregivers’ burden in people caring for a person with PD.

Methods

We analyzed the data set from a multinational online survey the Parkinson's real‐world impact assesSMent (PRISM) focusing on medication use, comorbidities, health‐related quality of life, relationship changes and the use of healthcare and supportive care resources by people with PD and their carers. Structured questionnaires including the Parkinson's disease quality of life questionnaire (PDQ‐39), non‐motor symptoms questionnaire (NMSQuest) and the Questionnaire for impulsive‐compulsive disorder in Parkinson's disease (QUIP) were applied. Caregiver burden was assessed by the Zarit Burden Interview (ZBI).

Results

In a cohort of 245 dyads (patient and respective caregiver), caregivers reported a mild to moderate burden. Carers’ perception of PD impact in partnership, financial burden, hours of care, patient's age, hypersexuality and health‐related quality of life (HRQoL) were found to be significant contributory factors to caregiver burden. Taken together these variables explained 66.8% of the variance in the Interpretation of the ZBI total score.

Conclusions

Caring for a person with PD entails substantial burden, particularly when the caregiver perceives greater changes in partnership dynamics, dedicates more time to caregiving tasks, has financial burden, and when the patient is older, reports worst HRQoL and has sexual compulsive urges.

Parkinson's disease (PD) is a common neurodegenerative movement disorder. ¹ As a chronic, progressive, and complex disease manifested by a broad spectrum of motor and non‐motor features, ² PD greatly impacts an individual's functioning, often leading to major disabilities and dependencies. ³ Therefore, people with PD (PwP) may need assistance with care, medication compliance, prompting for self‐care, social involvement, and personal safety. ⁴ , ⁵ These many responsibilities are usually assumed by an informal caregiver ⁶ who provides physical, social and emotional support for years or decades of PD progression, delaying or avoiding PwP institutionalization. ⁴ , ⁷ This chronic strain leaves caregivers more vulnerable to burden, especially when lacking adequate training and support. ⁶ , ⁸

The caregiver burden (CB) is defined as a multidimensional concept that expresses the negative physical, emotional, social, spiritual, and financial state induced by the demands of providing care to a person with an illness or a disability. ⁹

Prior studies suggest that nonmotor symptoms exert a stronger influence on CB compared to motor symptoms ⁴ ; and that CB increases as the person with PD becomes more dependent on activities of daily living ¹⁰ and patient's quality of life worsens. ¹¹ Directly related to the caregiver, the risk factors reported as affecting CB include being female, having a lower level of education, co‐habitation, providing care for long hours, having depression, being socially isolated, being under financial stress, and having no choice but to be a caregiver. ¹² Specifically in PD, few studies concentrated on the identification of caregivers’ characteristics that could predict their burden ¹⁰ , ¹¹ , ¹³ , ¹⁴ , ¹⁵ , ¹⁶ , ¹⁷ and yielded mostly inconsistent results. Only caregiver mood disorders seem to routinely correlate to higher burden. ⁴

Therefore, the CB‐associated factors in PD are still not completely clarified. By using a large and multicenter sample, this study aims to identify the main contributing factors to the burden of PD patients’ caregivers.

Methods

Data used in this study were obtained from the Parkinson's real‐world impact assesSMent (PRISM) database (www.prism.bial.com) (accessed on November 11, 2021).

PRISM Study Design and Population

The PRISM study protocol have been previously published. ¹⁸ Briefly, PRISM was an observational, multinational, cross‐sectional survey developed by an international scientific committee in collaboration with the cure Parkinson's trust. The study was conducted in six European countries—France, Germany, Italy, Portugal, Spain, and the United Kingdom—and data were collected using an online survey comprising two main sections: the first was answered from the perspective of the person with PD, and the second from the perspective of the main caregiver.

Participants were recruited with help from PD advocacy groups and specialist PD clinics, through email, social media campaigns and leaflets, aiming to reach the maximum patient‐caregiver dyads in each target country. Healthcare professionals had no direct role in recruitment and contribution to this study was voluntary, allowing omission to questions that the respondent did not want to answer. Before entering the survey, participants were informed that all information would be treated confidentially and stored securely, in accordance with General Data Protection Regulation. ¹⁹

Assessment

It was advised that, whenever possible, the survey should be completed by PwP and caregiver together.

In the section of the questionnaire addressed to PwP was collected socio‐demographic data and information on pharmaceutical treatment, comorbidities, the utilization of healthcare and supportive care resources, and the impact of PD on employment and relationships. Health‐related quality of life (HRQoL) of PwP was evaluated by the Parkinson's disease quality of life questionnaire (PDQ‐39; Oxford University Innovation Limited) ²⁰ , ²¹ ; the prevalence of non‐motor symptoms was deemed using the non‐motor symptoms questionnaire (NMSQuest; International Parkinson and Movement Disorder Society, Inc.) ²² ; and impulsivity analysis was based on the questionnaire for impulsive‐compulsive disorder in Parkinson's disease (QUIP). ²³ Questions relating sexual relationships were placed at the end of the survey and taken from the medical outcomes study sexual functioning scale. ²⁴

The carer questionnaire section allowed to obtain socio‐demographic data and information on the weekly hours spent caring for PwP, the utilization of social networks and supportive care resources to help them with this task and the impact of PD on partnership.

The CB was evaluated through the Zarit burden interview (ZBI; Mapi Research Trust). ²⁵ It is a questionnaire constituted by 22 questions that reflects the impact of the patient's disabilities on various aspects of carer's life, including physical, emotional, social and financial domains. Each item can be scored from 0 to 4: 0 for “never,” 1 for “rarely,” 2 for “sometimes,” 3 for “often,” and 4 for “always/nearly always.” ²⁵ The total score ranges from 0 to 88 and can be categorized for interpretation as follows: 0–20 corresponds to little or no burden; 21–40 represents mild to moderate burden; 41–60, moderate to severe burden; and 61–88, severe burden. ²⁶ The ZBI has been validated for use in all of the languages used in PRISM. ¹⁸

Data Analysis

The primary analysis of the study was the assessment of the key CB‐contributing factors, evaluated by the Zarit burden interview. On that account, we only included in the sample the responses of PwP whose caregiver answered completely to the ZBI (questions 1 to 22).

Statistical Analysis

Statistical analysis was performed using Microsoft Excel (Microsoft, Seattle, WC, USA) and SPSS for Windows, version 26.0 (SPSS Inc., Chicago, IL, USA).

Descriptive statistics summarized mean, standard deviation (SD), median, interquartile range (IQR), minimum and maximum for continuous variables; and categorical variables were depicted with frequencies and percentages. For operational reasons, the answers “prefer not to say,” “don't know,” and “not applicable,” were assumed as missing data.

Individual associations between patient‐and caregiver‐related variables (having <10% of missing data) with the dependent categorical variable—Interpretation of Zarit Burden Interview score—were examined resorting univariate ordinal regression, for possible inclusion in the subsequent stepwise multiple regression model. The level of statistical significance was set at 0.05. After excluding variables with high collinearity to ZBI, the remaining variables were selected for the multiple regression analysis to determine the best predictors of CB in PD. In the final multivariate model, the Durbin–Watson statistic was used to assess autocorrelation and was found to be 1.915. Tolerance (0.345–0.932) and variance inflation factor (1.073–2.895) values were both within acceptable limits.

The association between CB and financial burden was assessed using the Interpretation of ZBI total score as the dependent variable and item 15 of ZBI (“Do you feel that you don't have enough money to take care of your relative in addition to the rest of your expenses?”) as the independent variable. For this reason, ZBI's item 15 was converted to a binary variable, where the answer “Never” was codified as “no financial burden” and answers “rarely” “sometimes,” “often” and “always/nearly always” were considered “financial burden”. Collinearity was examined to determine the viability of this analysis and financial burden was considered as a binary independent variable for both univariate and multiple regressions.

Results

General Data

A total of 245 patient‐caregiver dyads were studied. The sociodemographic and clinical characteristics of patients, including information on healthcare services utilization, were shown in Table 1. Details on the burden of disease were depicted in Table 2, and the utilization of community services and existence of out‐of‐pocket costs were illustrated in Figures 1 and 2, respectively. In the studied PwP population, 58.4% were male, the mean (SD) age at the beginning of survey was 68.9 (9.9) and median (IQR) disease duration was 7 years. ⁴ , ⁵ , ⁶ , ⁷ , ⁸ , ⁹ , ¹⁰ , ¹¹ , ¹² The majority of participants reported at least one comorbidity (68.6%) (Supplementary Table S1); one non‐motor symptom (99.2%) (full data not shown), being the mean (SD) NMSQuest score of 14.7 (6.0); and/or one impulse control behavior (49.8%) (Table 1). The median (IQR) PDQ‐39 summary index was 36.8 (21.6–51.0) (Table 2). Over 45% of participants mentioned expenses with aids and devices (66.9%), travels (51.3%), physiotherapy appointments (49.4%) and PD medication (46.0%) (Fig. 2). In the caregivers’ group, 64.3% were female and 51% were aged over 65 years old (Table 3). In 80.4% of cases the caregiver was the partner/spouse of the person with PD (Table 3). Further information on caregivers’ characteristics can be seen in Table 3 and a descriptive sub‐analysis of the burden of care according to the type of relationship to PwP and caregiver's gender was summarized in Supplementary Table S2.

TABLE 1.

Sociodemographic and clinical characteristics of people with Parkinson's disease

Characteristic	Statistic
Country, n (%)
Italy	79 (32.2)
United Kingdom	70 (28.6)
Spain	35 (14.3)
Portugal	30 (12.2)
France	22 (9.0)
Germany	9 (3.7)
Gender (male), n (%)	143 (58.4)
Age at beginning of survey, y, mean (SD)	68.9 (9.9)
Age at diagnosis, y, Mean (SD)	60.4 (11.5)
Time since PD diagnosis, y, median (IQR)	7.0 (4–12)
Distance from home to a specialist, n (%)
<30 miles/50 Km	204 (84.6)
30–60 miles/50–100 Km	21 (8.7)
>60 miles/100 Km	16 (6.6)
Level of education (highest), n (%)
Primary	63 (26.3)
Secondary	36 (15.0)
Tertiary	141 (58.8)
Number of present anti‐PD treatment classes, median (IQR)	2 (1–3)
Under levodopa‐containing therapy, n (%)	212 (88.7)
Under dopamine‐agonists containing therapy, n (%)	112 (47.3)
Number of comorbidities per person, median (IQR)	1 (0–2)
Non‐motor symptoms questionnaire score, mean (SD)	14.7 (6.0)
Impulse control behaviors, n (%)
At least one	119 (48.9)
Binge eating	55 (22.5)
Compulsive shopping	39 (16.1)
Hypersexuality	32 (13.6)
Hobbyism	29 (12.0)
PD medication abuse	26 (10.7)
Gambling	11 (4.6)
Appointment(s) (yes), n (%)
Specialist	232 (95.1)
General Practice	224 (91.7)
Physiotherapy	161 (66.3)
Primary care nurse	88 (36.1)
Specialist PD nurse	66 (27.0)
Mental health	53 (22.2)
Speech/Language therapist	49 (20.2)
Occupational therapist	36 (14.9)
A&E department admission(s) (yes), n (%)	82 (34.0)
Hospital admission(s) (yes), n (%)	52 (21.6)

Abbreviations: A&E, accident & emergency; IQR, interquartile range; PD, Parkinson's disease; SD, standard deviation.

TABLE 2.

The burden of disease of people with Parkinson's disease

Characteristic	Statistic
Paid employment (N = 173), n (%)
Yes	14 (8.1)
No	150 (86.7)
Other (eg, sick leave or career break)	9 (3.7)
Early retirement (N = 198), n (%)
Retired early	63 (31.8)
Mainly due to PD	47 (23.7)
Not mainly due to PD	16 (8.1)
No early retirement	135 (68.2)
Degree of PD impact on family relationships, n (%)
Not at all	27 (11.3)
Slightly	45 (18.9)
Moderately	67 (28.2)
Very much	70 (29.4)
Extremely	29 (12.2)
Degree of impact of PD progression on family relationships, n (%)
Not at all	34 (14.7)
Slightly	56 (24.1)
Moderately	61 (26.3)
Very much	55 (23.7)
Extremely	26 (11.2)
PDQ‐39 summary index, median (IQR)	36.8 (21.6–51.0)

Abbreviations: IQR, interquartile range; PD, Parkinson's disease; PDQ‐39, Parkinson's disease quality of life questionnaire.

Figure 1.

Supportive care services utilization in the last 12 months by People with Parkinson's Disease (<10% missing values).

Figure 2.

Existence of expenditures related to Parkinson's disease (PD). When the number of people with PD (PwP) answering had >10% missing values compared to the total number of participants (N = 245), N is indicated in parenthesis.

TABLE 3.

Characteristics of caregivers and burden of care

Characteristic	Statistic
Gender (female), n (%)	157 (64.3)
Age group, y, n (%)
Less than 45 years	19 (7.8)
45 to 64 years	101 (41.2)
65 to 84 years	123 (50.2)
85 years and over	2 (0.8)
Relationship of caregiver to person with PD, n (%)
Partner/spouse	197 (80.4)
Sibling	35 (14.3)
Parent	8 (3.3)
Friend	3 (1.2)
Daughter/Son	2 (0.8)
Weekly hours of care provided by the main caregiver in the last 3 months, h
Median (IQR)	17.0 (5.0–50.0)
0 to 7, n (%)	74 (31.2)
8 to 28, n (%)	72 (30.4)
29 to 49, n (%)	31 (13.1)
50 to 100, n (%)	30 (12.7)
Above 100, n (%)	30 (12.7)
Existence of other caregivers of the PwP, n (%)
Yes	106 (44.2)
Family (N = 106)	73 (68.9)
Friends (N = 105)	32 (30.5)
Paid caregiver a (N = 104)	29 (27.9)
Paid nurse (N = 105)	9 (8.6)
Degree of PD impact on family relationships, n (%)
Not at all	67 (28.2)
Slightly	58 (24.4)
Moderately	62 (26.1)
Very much	39 (16.4)
Extremely	12 (5.0)
Degree of impact of PD progression on family relationships, n (%)
Not at all	58 (24.5)
Slightly	62 (26.2)
Moderately	56 (23.6)
Very much	52 (21.9)
Extremely	9 (3.8)
Sexual Relationship affected by illness (N = 168)
Yes, n (%)	100 (59.5)
ZBI total score (0–88)
Median (IQR)	25 (11.5–39.0)
Mean (SD)	17 (27.6)
Minimum	0
Maximum	83
Interpretation of ZBI total score, n (%)
Little/no burden (0–20)	106 (43.3)
Mild/Moderate burden (21–40)	84 (34.3)
Moderate/severe burden (41–60)	48 (19.6)
Severe burden (61–88)	7 (2.9)

^a Paid caregiver (non‐nurse).

Abbreviations: IQR, interquartile range; SD, standard deviation; PwP, people with Parkinson's disease; PD, Parkinson's disease; ZBI, Zarit burden interview.

Correlates and Determinants of Caregiver Burden

In univariate ordinal regression, 28 out of the 56 patient‐related variables and 7 out of 8 caregiver‐related variables showed significant correlations to ZBI score interpretation, as discriminated in Supplementary Tables S3–S8. Despite having significant correlations, some patient‐related variables such as discriminated comorbidities and PDQ‐39 domain score were not considered to enter stepwise multiple regression to avoid repeated information with total number of comorbidities and PDQ‐39 Summary Index, respectively.

To identify the best predictors of CB, a stepwise multiple regression analysis was performed, entering as independent variables those that showed a significant association in the simple ordinal regression with Interpretation of ZBI score, after excluding collinearity between them with both tolerance (>0.3) and the variance inflation factor (<3). Starting with 22 variables (15 patient‐related and 7 caregiver‐related) (Supplementary Table S9) that might theoretically be good predictors of CB, results presented a model comprising caregiver's perception on impact of PD progression in family relationships, financial burden, PDQ‐39 Summary Index, hours of care provided by main carer, impact of PD in family relationships perceived by caregiver, patient's age and hypersexuality, that could explain 66.8% of the total variance of the ZBI level (R ² = 0.668, P < 0.0001) (Table 4).

TABLE 4.

Caregiver burden associated factors according to results of ordinal multiple regression

Variable	Unstandardized coefficient		Standardized coefficient	t	P‐value
	B	SE	Beta (β)
Constant	−1.667	0.429		−3.887	<0.0001
Impact of PD progression in family relationships a	0.247	0.078	0.23	3.159	0.002
Financial burden	0.661	0.133	0.251	4.98	<0.0001
PDQ‐39 summary index	0.014	0.004	0.212	3.643	<0.0001
Hours of care provided by main caregiver	0.006	0.002	0.15	2.81	0.005
Age at beginning of survey b	0.02	0.006	0.147	3.096	0.002
Impact of PD in family relationships a	0.181	0.072	0.169	2.503	0.013
Hypersexuality b	0.452	0.188	0.107	2.409	0.017

^a From caregiver's perspective;

^b Of person with PD. R ² = 0.668, adjusted R ² = 0.655, F = 52.082.

Impact of Caring for People with PD

Hours of Care

Carers dedicated a median (IQR) of 17 (5.0–50.0) hours per week to caregiving tasks (ie, doctor appointments, shopping, cleaning, and driving) in the last 3 months. Around 44% confirmed receiving assistance from others, mainly from a family member (Table 3). The existence of other caregivers was more frequently reported by non‐spouse carers comparing to spouse carers (68.9% vs. 39.5%) (Supplementary Table S2).

Change in Relationships

The majority of carers (71.8%) stated that PD impacts their relationship with the patient slightly (24.4%), moderately (26.1%), very much (16.4%) and extremely (5.0%). Throughout PD progression 75.5% reported that this changed to slightly (26.2%) moderately (23.6%), very much (21.9%), and extremely (3.8%) (Table 3).

In Figure 3 were depicted the patient responses (horizontal‐colored bars) according to their matched caregiver's perspective on the questions: “Has your relationship with the person with Parkinson's suffered because of their illness?” (Fig. 3A), and “Do you feel that the impact of Parkinson's on your relationship has changed as the disease has progressed?” (Fig. 3B). Among the dyads whose caregiver answered “extremely” to the first question, 80% of patients were found to state the same degree of impact, 10% considered a moderate impact and 10% a slight impact (Fig. 3A). Carers (59.5%) also referred that their partner's illness affected their sexual relationship (Table 3).

Figure 3.

(A) Impact of Parkinson's disease (PD) on relationships for caregivers and matched person with PD and (B) Change in impact for carers and matched person with PD as disease progresses. Percentages of confluent perceptions are depicted in bold.

Zarit Burden Interview Results

The ZBI total score ranged from 0 to 83 (median, 25; IQR 11.5–39.0), corresponding to 43.3% of caregivers referring little/no burden, 34.3% mild/moderate burden, 19.6% moderate/severe burden and 2.9% severe burden (Table 3). Among the most frequently endorsed statements, more than half caregivers reported to, at least sometimes, being afraid of what the future holds for the patient (83.7%), feeling that the patient depends on them (72.7%), having difficulties to meet other responsibilities (ie, with family and work) (57.9%), feeling that the person with PD seems to expect them to be the only person they could depend on (54.7%) and not having enough time for themselves (51.4%). To the question “Do you feel that you don't have enough money to take care of your relative in addition to the rest of your expenses?” (ZBI15) 53.5% responded “never,” 12.7% “rarely,” 15.1% “sometimes,” 12.2% “often,” and 6.5% “always/nearly always.” Details on caregiver's answers to ZBI can be seen in Figure 4.

Figure 4.

Caregivers’ response to each question of the Zarit burden interview (ZBI), ordered 1–22. All participants completed the questionnaire (N = 245).

Discussion

Our study found a mild to moderate burden which is in line with other studies whose sample sociodemographic characteristics were similar to ours. ¹¹ , ¹³ Our multiple regression model (Table 4) highlighted caregiver's perception of a greater illness impact in partnership, financial burden, longer hours of care, patient's older age, higher PDQ‐39 summary index and hypersexuality as the best significant contributors to CB in PD. Several other candidate CB‐associated factors were found to be significant in univariate analysis (Supplementary Tables S3–S8), even though exclusion from final model, suggesting that their associations with caregiver burden were smaller or might be partially mediated by other factors.

Caregiver‐Related Determinants of Caregiver Burden

Caregiver's judgment about the impact of illness on the relationship played a key role in their burden. Carers, who were in the great majority (80.4%) the spouse or life partner of the person with PD, asserted that illness impacted family relationships (71.8% to 75.5%) and adversely affected their sexual relationship (59.5%). According to our sub‐analysis based on caregiver's gender and type of relationship to PwP (Supplementary Table S2), female caregivers and non‐spouses reported an at least moderate impact on relationship more frequently (53.9% to 53.6% vs. 36.5% to 42.2%, and 48.9% to 58.7% vs. 47.4%, respectively). Considering each subgroup individually, we found that only in female spouses’ subgroup occurred a decrease in the percentage of caregivers reporting a moderate to extreme impact when the progressive nature of PD was highlighted (54.3% to 53%). Further analysis could elucidate the factors contributing to partnership changes due to PD considering its progression.

In our multivariate model, financial burden also exerted a substantial influence in general burden associated with caregiving. Around 67% had out‐of‐pocket costs with aids and devices, 51.3% invested on travels associated with PD and 46% had expenditures with PD medication. Among healthcare services, the ones that most frequently required expenditures were physiotherapy (49.4%) and speech therapist appointments (42.4%). Regarding supportive care services, participants reported having expenses mainly with overnight assistance (18.7%) and support at social events (12.3%) (Fig. 2). Furthermore, participants reported compulsive shopping (16.1%) and gambling (4.6%) (Table 1). In PRISM, neither employment status nor income of the caregivers were assessed, but prior studies indicated that hidden costs ¹⁴ and low monthly income ²⁷ were associated with experiencing greater burden, whereas “outside employment” correlated to lower levels of burden. ¹⁰ Bekdemir et al. ¹⁰ suggested that the relationship between employment and less burden stemmed from the inability to be a full‐time caregiver and, consequently, have shorter caregiving time.

In our study, time investment (objective burden) revealed to be determinant of higher CB (subjective burden). Caregivers reported spending a median (IQR) of 17 (5–50) h per week assisting the patient's needs, including doctor appointments, shopping, and cleaning. Sparse studies evaluated the direct association between hours of care with ZBI, ²⁸ , ²⁹ not exclusively in advanced stages of PD, ¹¹ , ¹⁷ allowing comparison to our results. P. Martinez‐Martin et al ²⁸ denoted a positive moderate‐strong correlation in univariate analysis but caregiving time was not among the major CB‐associated factors. In the same study, PD clinical aspects (ie, H&Y stage and dependency in activities of daily living) were the most important factors, suggesting that these exert a stronger influence in caregiving demands, being more burdensome than the factor time itself. In fact, some reports considered functional dependency in daily activities as a key CB‐associated factor. ¹⁰ , ¹⁵

Patient‐Related Determinants of Caregiver Burden

Regarding CB‐associated factors directly related to PwP, three were considered significant in multiple regression—age, HRQoL (measured by PDQ‐39) and hypersexuality.

In our sample, PwP were aged 68.9 ± 9.9 years and reported a wide range of co‐morbidities (Supplementary Table S1). Despite PD duration, comorbidities tend to compound the process of aging, converging to poorer physical health and amplifying caregiver burden. ³⁰ , ³¹ In our study, due to collinearity with the total number of comorbidities, the detailed ones were excluded from multivariate analysis and this should be taken into consideration when interpreting age impact on CB, since, in univariate analysis, stroke, dementia, peripheral vascular disease, depression, and heart issues had significant (P < 0.05) and strong (r ≥ 0.6) correlations with ZBI (Supplementary Table S4). In literature, demographics were shown not to be major CB‐associated factors, whereas depression and cognitive impairment had inconsistent results. ³²

The identification of HRQoL as a main CB‐associated factor reflects the core importance of patient's perspective over their global well‐being to caregiver, being in line with previous reports. ³³

Impulse Control Disorders, often associated with dopaminergic therapy, ³⁴ were previously associated to higher burden when considered as a whole. ³⁵ To our knowledge, this is the first study to evaluate the association between individual ICDs and CB. Hypersexuality was the only one to significantly correlate to ZBI in univariate analysis (Supplementary Table S4) and was also considered as a key CB‐associated factor (Table 4). This compulsive sexual behavior, encompassing qualitative and quantitative changes in sexual habits, ³⁶ was present in 13.6% PwP in our sample—a concordant prevalence to the one found when a self‐completed questionnaire (QUIP) is used to assess ICDs. ³⁷ Of note, patients usually underestimate the presence and severity of ICDs, particularly hypersexuality, ³⁸ and A.H. Erga et al ³⁹ suggested that only caregiver‐reported ICDs negatively impact the burden of the carer. Therefore, clinicians should not neglect the importance of monitoring these behaviors with both patient and caregiver, to avoid and manage adverse consequences in partnership.

Other Findings

In general, PwP tended to perceive a greater impact of the disease at relationship's quality, comparing to their caregiver (Tables 2 and 3) and it was showed that this perception is not always concordant between caregiver and respective care recipient (Fig. 3). Ricciardi et al ⁴⁰ considered that, in chronic diseases, caregivers could handle well the task of caregiving but they might lack the skills to relate as partners. This could be a possible explanation to our results: caregivers might focus on symptoms’ control reframing it as emotional commitment and take the new role as inherent to a quality relationship; while the person with PD might face himself/herself as the burden and the disease as guilty of the change in caregivers’ lifestyle, leading to higher perceived relationship disruption. Notwithstanding, given that in PRISM it was advised that the questionnaire should be completed together and that the patient's responses preceded those of the caregiver, the second may have underestimated the impact of the disease on the relationship.

Considering the association between caregiving hours with ZBI, we could hypothesize that the existence of multiple informal caregivers (family, friends and/or acquaintances) could contribute to CB reduction. Nevertheless, in our sample, it was not a significant protector of burden in univariate analysis (Supplementary Table S8). A review on Caregiver Burden in Multiple Sclerosis ⁴¹ highlighted the need for community support outside carers’ usual environment (ie, family and hospital staff) to lessen their burnout. Therefore, instead of focusing on replacing the main carer or reducing their time with the patient, attention should be taken to their self‐efficiency perception and we should strive to help them develop better coping strategies to handle the changeful daily reality.

Despite PD being considered a primarily motor disorder, a wide range of non‐motor symptoms are reported by PwP across all stages of PD. Similarly to previous reports, ⁴² , ⁴³ in our sample, patients had on average 14.7 (6.0) non‐motor‐symptoms. We failed to demonstrate NMSQuest total score as one of the main CB‐associated factors. Conversely, Kalampokini et al ¹⁷ found the total non‐motor symptoms scale (NMSS) score to be among the determinants of CB in a large European sample. These contrasting results could be explained by the different measurements used: while NMSQuest (used in PRISM) is a dichotomic questionnaire to screen symptom frequencies, NMSS (used in CLaSP) also assesses severity. ⁴⁴ Moreover, non‐motor symptoms are known to associate with worst HRQoL, ⁴⁵ , ⁴⁶ , ⁴⁷ , ⁴⁸ thus, in our analysis, their impact on CB might have been mediated by PDQ‐39 summary index.

Limitations

This study has some limitations that must be noticed. A drawback of the observational nature of the study is the risk of bias and confounding. In fact, we cannot exclude that the results and drawn conclusions were biased toward younger and higher educated patients, and caregivers most actively involved in the caring of PwP. Moreover, our analysis was restricted to the data provided by the PRISM, which questionnaire did not include validated markers of disease severity and stage, disability level, non‐motor symptoms severity, nor information on the local of residence (ie, living at home versus in long‐term care facilities) and on levodopa‐equivalent daily dose. These aspects, along with the relatively low disease duration in our sample, increase the chances of missing contributory factors to CB related to disease progression. Also, the lack of a formal supervision and the self‐completion of the survey has an impact on the discussion of the prevalence of comorbidities, NMS and ICDs. As previously outlined, we acknowledge that some caregiver‐related variables, such as socioeconomic status, health problems (eg, anxiety and depression), personality traits, caregiving experience and time of caring (ie, years) were not assessed and might have a significant impact on burden perception. Finally, the cross‐sectional design of the study does not allow conclusions on the direction of causality and limits the generalization of results to similar samples.

Implications for Clinical Practice and Future Research

There is sparse robust evidence on strategies to alter the trajectory of CB. Some suggestions under consideration include self‐management support for PwP and caregivers, and optimizing symptom control through medication and psychotherapy, ⁴ , ⁶ in the scope of a home‐based community‐centered integrated care. ⁶ Health technology and flourishing disease‐modifying therapies ⁴⁹ might also contribute to mitigate carer distress. Our comprehensive study, with a large and multicenter sample, could help shape these interventions by adding to current knowledge on CB major contributory factors. Even though, in the setting of a progressive neurological disease with fluctuating symptoms, longitudinal studies considering other important variables, are needed to construct more accurate models of CB‐associated factors to better guide the assessment of the complex system patient‐carer. Henceforward, research projects could evaluate the impact of the previously mentioned interventions in the key CB‐contributing factors, stimulating the improvement and development of specific therapies to relief and prevent this multifactorial distress.

Conclusion

In summary, caregivers in this sample reported a mild to moderate level of burden, mostly determined by changes in partnership due to PD, especially as disease progressed, financial burden, longer time per week spent caring, patient's age, HRQoL and sexual compulsive urges.

Author Roles

(1) Research Project: A. Conception, B. Organization, C. Execution; (2) Statistical Analysis: A. Design, B. Execution, C. Review and Critique; (3) Manuscript Preparation: A. Writing of the First Draft, B. Review and Critique.

G.M.S.: 1A, 1B, 1C, 2A, 2B, 2C, 3A

R.B.M.: 1A, 1B, 2A, 2C, 3B

D.A.: 2A, 2C

J.J.F.: 1A, 1B, 2C, 3B

Disclosures

Ethical Compliance Statement: We confirm that we have read the Journal's position on issues involved in ethical publication and affirm that this work is consistent with those guidelines. A process was undertaken in the United Kingdom to determine whether ethical approval was required for the study, using online tools provided by the NHS England Health Research Authority. This indicated that the study was research (“Is my study research?” http://www.hra-decisiontools.org.uk/research/) but did not require NHS Research Ethics Committee (REC) approval (“Do I need NHS REC approval?” http://www.hra-decisiontools.org.uk/ethics/). Informed consent was obtained from all participants.

Funding Sources and Conflicts of Interest: PRISM study and database was funded by BIAL–Portela & Cª, S.A., designed in collaboration with The Cure Parkinson's Trust, an advocacy group based in the United Kingdom (UK), and reviewed by the PRISM steering committee. The authors declare no external sources of funding and no conflicts of interest.

Financial Disclosures for the Previous 12 Months: G.M.S., R.B.M and D.A. declare no additional disclosures to report. J.J.F received grants from GlaxoSmithKline, Grunenthal, Fundação MSD (Portugal), TEVA, MSD, Allergan, Novartis, Medtronic and Angelini. Received consultancy and speaker fees and participated in advisory boards for GlaxoSmithKline, Novartis, TEVA, Lundbeck, Solvay, BIAL, Merck‐Serono, Merz, Ipsen, Biogen, Acadia, Allergan, Abbvie, Sunovion Pharmaceuticals, Zambon, Affiris, Roche, ONO and SK Chemicals.

Supporting information

Table S1. Prevalence of comorbidities in people with Parkinson's disease.

Table S2. Burden of care according to relationship to person with PD and caregiver's gender.

Table S3. Association between patient's sociodemographic characteristics and caregiver burden–univariate ordinal regression results.

Table S4. Association between patient's clinical characteristics and caregiver burden–univariate ordinal regression results.

Table S5. Association between patient's utilization of healthcare services in the last 12 months and caregiver burden–univariate ordinal regression results.

Table S6. Association between patient's use of supportive care resources in the last 3 months and caregiver burden–univariate ordinal regression results.

Table S7. Association of PDQ‐39 score and patient's perspective on PD impact in relationships with caregiver burden–univariate ordinal regression results.

Table S8. Association between caregiver characteristics and caregiver burden–univariate ordinal regression results.

Table S9. Variables selected for the multiple regression analysis.